Bug 10709

Created attachment 9679 [details]
stex3d, sphere

Created attachment 9680 [details]
stex3d, torus

Probably a bug with swrast fallbacks or mesa core (as the driver doesn't implement 3d textures). Not sure what would have caused this.

This bug still happens with the newly-released Mesa 6.5.3.

Created attachment 10030 [details]
seom video capture file of stex3d output

I thought this problem might be clearer if I provided a moving picture. This file was created using yukon:

http://www.neopsis.com/projects/yukon/

The seom SVN repository includes the seom-player application to view this file.

One odd thing I noticed here: when I do view this file, the playback window turns completely black unless I hold the <Alt><TAB> keys down. But so long as I do that, I can see the playback. I have also seen this exact same "blank window" behaviour while trying to play "Deus Ex" under Wine in OpenGL mode, so I'm thinking it's Mesa/R100 related rather than a problem with seom-player.

I have just tried playing this .seom file with my R200 (rv280?). The playback window still turns black until I press <Alt><TAB>, although it then continues to play normally when I release them.

I can't reproduce this. I get similar breakage though with hyperz enabled, which is expected and nothing new (reading/writing z buffer directly, which happens with swrast fallbacks, simply does not work at this point, and is basically the reason hyperz is not enabled by default).

(In reply to comment #7)
> I get similar breakage though with hyperz enabled, which is expected and nothing
> new.

My seom-player behaves exactly the same with my R200, regardless of whether HyperZ is enabled or not. (I am using git Mesa with Xorg packages from FC6.)

Section "Device"
        Identifier  "Radeon 9200"
        Driver      "radeon"
        VendorName  "Sapphire Tech"
        BoardName   "ATI Technologies Inc RV280 [Radeon 9200]"
        Option      "AGPMode" "8"
        Option      "GARTSize" "128"
        Option      "AccelMethod" "EXA"
        Option      "AccelDFS" "on"
        Option      "EnablePageFlip" "on"
        Option      "RenderAccel" "on"
EndSection

I should probably raise this issue as a separate bug, but anyway - here is my .drirc file for my rv280 card. And I'll reiterate that seom-player behaves the same regardless of whether hyperz is "true" or "false":

<driconf>
    <device screen="0" driver="r200">
        <application name="Default">
            <option name="force_s3tc_enable" value="false" />
            <option name="no_rast" value="false" />
            <option name="fthrottle_mode" value="2" />
            <option name="tcl_mode" value="1" />
            <option name="texture_depth" value="1" />
            <option name="def_max_anisotropy" value="1.0" />
            <option name="no_neg_lod_bias" value="false" />
            <option name="texture_units" value="6" />
            <option name="dither_mode" value="0" />
            <option name="hyperz" value="true" />
            <option name="round_mode" value="0" />
            <option name="texture_blend_quality" value="1.0" />
            <option name="nv_vertex_program" value="false" />
            <option name="color_reduction" value="1" />
            <option name="vblank_mode" value="1" />
            <option name="allow_large_textures" value="2" />
        </application>
    </device>
</driconf>

Since seom-player uses the Xv extension, I tried experimenting with gxvattr on my rv280 machine. The problem does not happen if I enable XV_ALPHA_MODE, and starts again when I disable it.

I will test whether this also fixes things for my R100.

(In reply to comment #8)
> (In reply to comment #7)
> > I get similar breakage though with hyperz enabled, which is expected and nothing
> > new.
> 
> My seom-player behaves exactly the same with my R200, regardless of whether
> HyperZ is enabled or not. (I am using git Mesa with Xorg packages from FC6.)

I was commenting on the original issue not the one with the seom-player - I know nothing of this though yes it is unlikely to be affected by hyperz...

Created attachment 10038 [details]
BZ2-compressed seom video capture of stex3d; HyperZ disabled

Disabling HyperZ does affect the stex3d results, but it is still not correct.

As an aside, setting XV_ALPHA_MODE=1 "fixes" seom-player on both R100 and R200. I will raise a separate bug for this.

Retested after bug #13080 was fixed: Disabling HyperZ now allows the stex3d demo to run properly.

*** This bug has been marked as a duplicate of bug 13080 ***

Mass version move, cvs -> git